5i 
Trigonocarpus Shorensis , sp. nov . 
heads the column. In Section f the sarcotesta is much broken, hence the 
figure for g is probably double its real value. If this be the case, then the 
number of additional sacs appearing at any given level is about twenty-one, 
or sixty-three for the whole circumference. The whole interval included 
between S. 31, b and S. 31, i is about 14 mm., and out of the total of 195 
separate secretory elements encountered, not more than eight persist 
throughout. A few are only present in one section, whilst the greater 
number extend into the next (an interval between sections of about 2 mm.) 
with a falling off in each successive section. 
S. 31. 
b 
c 
d 
e 
/ 
_g 
h 
i 
Total. 
Section S. 31. c 
46 
23 
69 
„ d 
45 
15 
26 
86 
„ * 
35 
12 
24 
22 
93 
„ / 
2 
2 
2 
2 
10 + 
2 
» g 
23 
6 
2 
8 
10 
37 
86 
„ h 
20 
5 
2 
7 
6 
32 
11 
8 3 
l» * 
8 
4 
2 
6 
6 
13 
5 
20 
64 
The level of k in the Manchester series is the base of the micropyle, 
and of c that at which the nucellus is attached to the chalazal cushion. 
The number of secretory sacs in each case represents approximately one- 
third of the total at that level. 
Man. Coll, c 34 
,, d 46 
v e 65 
„ / 82 
» g 9 8 
Man. Coll, h 101 
,, i 92 
» J 93 
„ k 109 
A study of the two series shows that, in spite of irregularities, there 
was a decrease in number both towards the apex and base. From a com- 
parison of the same sacs cut at different levels, they are seen to have tapered 
gradually both above and below, so that where cut across near their termina- 
tions they are of very small size and, where the surrounding wall is disclosed, 
very closely resemble developmental stages in the formation of secretory 
canals, the surrounding sheath simulating an epithelium, which may perhaps 
be the explanation of the similar appearances described by Professor Seward 
in the medullosean petiole Rachiopteris Williamsoni} 
A single secretory element, as seen in one section, is in several instances 
replaced by two in intimate contact in the succeeding section (PI. V, Fig. 13) ; 
the fact that both are usually of nearly the same cross-section seems to favour 
the view that this is due to branching. 
The normal course followed by the secretory elements was slightly 
sinuous and approximately parallel to the direction of the surface, with an 
outward tendency exhibited by the more interior. The distribution was 
irregular, with a marked increase — especially of the smaller — towards the 
periphery ; whilst the largest (0-124-0-166 mm.) are found midway between 
the sclerotesta and the external surface (cf. Text-fig. 3, p. 42). 
1 Ann. Bot., vol. viii, No. xxx, 1894. 
